Skip to content

Features

What Mandrel does today—with an explicit focus on making operational data easier to live with, including assistant AI your org can enable. If it is not listed, we have not shipped it yet.

Collections & structure

Create datasets inside your organization that look like the things you actually track—customers, assets, requests—not anonymous rows waiting for a legend.

  • Fields can enforce format and requiredness so bad data trips before it spreads.
  • Parent/child collections when your model is naturally hierarchical.
  • Soft delete and restore when you want undo, not panic.

Forms that people can follow

Break long intake into sections so the form reads like a conversation, not a wall of boxes.

  • Sectioned layouts for scanning and training new teammates faster.
  • Templates you can repeat so the hundredth entry matches the first.

Users, roles & who-sees-what

Map your org to clear rules: who can read, create, update, or delete in each collection—without handing everyone master keys.

  • Per-collection permission profiles for sharing that matches reality.
  • Super-administration paths for tenant-wide controls when that mode is available.

Records, search & history

Work with structured rows in predictable lists, and give admins a trail when someone needs to reconstruct a decision.

  • Search, sort, and page through large sets without exporting first.
  • Audit listings for administrators reviewing who did what.

Automations

When your deployment allows it, tie actions to record events—so the boring follow-ups happen on their own, with a log when you need proof.

  • Execution history stays visible when automation platforms are enabled.

AI assistant & in-product help

Mandrel treats AI as part of the product surface—not a side gimmick—so contributors can draft, summarize, and get unstuck against real collections, within org context and sensible limits. Flip it on when your administrators are ready.

  • Built to reduce friction in everyday record work, not to bypass your policies.
  • Daily or org quotas may apply; subprocessors (e.g. inference providers) are disclosed.
  • Treat outputs as drafts—humans stay accountable for consequential decisions.

Building something that talks to Mandrel? Open the HTTP API reference — OpenAPI shapes, auth, and an explorer so you can poke endpoints without guessing. Prefer AI inside the product? See the assistant section above for how in-app help ties back to the same collections.